Iterative learning control with basis functions for media positioning in scanning inkjet printers

Abstract

Abstract-In printing systems, the positioning accuracy of the medium with respect to the print heads directly impacts print quality. In a regular document inkjet printer, the main task of the media positioning drive is to shift the medium after the printhead has finished a pass. Most media have the tendency to deform while it is being printed due to variations in temperature and moisture content. In order to improve print quality, we propose to move the medium during printing to counteract the deformation. These small scale trajectories are performed in an operating regime of which the dynamics considerably differ from the regular transportation step. Using iterative learning control with basis functions for both positioning tasks, the positioning accuracy of the drive is improved substantially; while keeping numerical cost low.

Using iterative learning control with basis functions to compensate medium deformation in a wide-format inkjet printer

“…ILC updates the command signal offline in a batch-to-batch fashion and hence naturally meets requirement R1. Suitable basis functions [33,4] are introduced to deal with the varying reference trajectories that are required for R2. The controller design is experimentally validated and compared with standard norm-optimal ILC in a reproducible experiment where a set of benchmark trajectories is used.…”

“…A potential approach to compensate the paper deformation is actively changing the longitudinal paper position during a lateral pass of the carriage [4]. The shape of the deformation gradually changes in time [35], and therefore the reference trajectory for the paper motion must also change each pass.…”

“…The shape of the deformation gradually changes in time [35], and therefore the reference trajectory for the paper motion must also change each pass. In [4], only measurements of the motor position were used for control. Despite the significant performance enhancement at the motor side, an evaluation at the paper revealed insufficient performance.…”

“…Although the approach in [4] is conceptually promising, the disturbances introduced by the medium position drive [16] cannot be suppressed by applying control using the measured motor position. Therefore, in the present paper, the use of a newly developed scanner in the control loop is proposed.…”

“…Potentially, this method achieves improved performance and extrapolation capabilities simultaneously. The earlier approach in [4] cannot be applied directly since the controller structure is not suitable for the additional non-real-time measurements, moreover, the extrapolation capabilities are not investigated. The results in [5] employ a more complex set of basis functions than considered here, to further increase performance and extrapolation capabilities.…”